Summary
Mike Kasper is a Principal Research Engineer and PhD candidate in Computer Science with 12 years of experience building perception and reconstruction systems for robots and large-scale production fleets. He blends deep academic research—differentiable path tracing, learned dense feature maps for relocalization, and light-source estimation—with hands-on productionization at Amazon and startups, optimizing PyTorch models with TensorRT/ONNX and deploying low-latency perception pipelines. At Durable he bridged program synthesis and retrieval-augmented generation, training GNN-guided retrieval and text-diffusion models and contributing to a pending US patent for AI-driven code synthesis. Based in Barcelona, he is equally fluent in CUDA/OptiX for real-time 3D rendering and in applying modern NLP techniques (LoRA, fine-tuned T5/BERT, Llama) to practical problems, a mix that uniquely positions him to translate novel perception research into scalable robotics products.
12 years of coding experience
11 years of employment as a software developer
Bachelor’s Degree Computer Science, Bachelor’s Degree Computer Science at Northern Arizona University
Doctor of Philosophy (Ph.D.) Computer Science, Doctor of Philosophy (Ph.D.) Computer Science at University of Colorado Boulder
English, Spanish